The mechanisms of human atrial fibrillation (AF) are poorly understood. Previous isolated animal heart[unreadable] experiments demonstrated that some cases of acute AF may be maintained by the uninterrupted periodic[unreadable] activity of a small number of discrete reentrant sites (rotors) located in the posterior LA wall, near the PV/LA[unreadable] junction. In those experiments, the fastest rotors acted as dominant frequency sources (drivers) that[unreadable] maintained the overall activity. This resulted in a hierarchy of local excitation frequencies throughout both[unreadable] atria. More recently, clinical studies have confirmed the existence of a hierarchical organization in the rate of[unreadable] activation of different regions in the atria of patients with paroxysmal and chronic atrial fibrillation. However,[unreadable] the mechanisms underlying such a hierarchical; distribution of frequencies in human AF has not been[unreadable] explored. Our general hypothesis is that both local activation frequency and degree of regularity, while[unreadable] different in different parts of the atrium, are distributed non-randomly, with different patterns of distribution in[unreadable] paroxysmal versus chronic AF patients. We further surmise that such patterns are the result of fibrillatory[unreadable] conduction of waves emanating from AF drivers localized at the site of highest frequency and organization[unreadable] activity, with a gradual reduction of activation frequency as the distance from the driver increases. Thus our[unreadable] Specific Aims are: 1. In patients with paroxysmal and chronic AF, to quantify online and with high resolution[unreadable] the dominant frequency (DF) and regularity index (Rl) of the endocardia! electrical signals as separate[unreadable] measures of rate and fragmentation, respectively. 2. Also in patients with paroxysmal and chronic AF, to[unreadable] differentiate between reentrant and focal AF drivers by studying the effects of adenosine infusion on the DF[unreadable] and Rl distributions. 3. In paroxysmal AF patients, to determine the "breakdown frequency" at which rapid[unreadable] pacing in the presence and the absence of adenosine results in wavefront fragmentation, reflected by a[unreadable] sudden change from 1:1 LA:RA activation to fibrillatory conduction. 4. In computer simulations, to study the[unreadable] mechanisms of initiation and maintenance of AF at the PV/LA junction using three different computer models[unreadable] with increasing anatomical complexity. Successful achievement of our specific aims should help us advance[unreadable] understanding of the mechanisms and manifestations of this complex arrhythmia and may help to directly[unreadable] improve the efficacy of pharmacological and ablative therapies in patients.